Design of face detection and recognition system monitor students during online examinations using Machine Learning algorithms

Aim:

          To ease the process of attendance during online classes and to prevent malpractice during online assessments.

Abstract:

           Today’s pandemic situation has changed the way of education to students. Education system had completely undertaken by online platforms. There are many platforms that provide services to schools, colleges and other educational intuitions. In addition to the online teaching, examination also has gone online. In this online education system, it is important to monitor the students presence i.e., attendance of the students, which plays vital role in making system. Even though educational institutions have adapted to online mode which is done through online portal, it is very challenging for staffs in taking attendance for a students. With the integrated webcam in online portals we can monitor the activities of students and malpractices done by them. But the attendance of the students is quite difficult and it is done manually by staffs by just seeing who are all in the streaming video. This can be bridled by face detection and recognition techniques using KNN algorithm. With this integration we can extract a facial feature vector which is otherwise known as embeddings and train it using Python Face Recognition Library, then identify the faces of the students before entering into examination dashboard. If it is matched the attendance system auto updates the presence of the particular student and we can also find the attendance percentage of the students, which is useful in analyzing the willing percentage of the students to write the exam online.

Existing System:

          In our education system, exam plays vital role in accessing and evaluating the student’s knowledge. These exams have several strict set of rules to be followed by the students in order to write their exams and these rules ensure that the student write their exams in proper way without malpracticing. Now as it’s a pandemic period this system is completely made online and the monitoring is done via webcam during the classes and exams. The educational institutions conduct exams for large number of students and it is difficult to monitor each one manually while some students take this as an advantage and started to impersonate someone like themselves during exams and scores better than the one who are true. To prevent this we are going to integrate a face recognition technique and machine learning algorithm in proposed system.

 Proposed System:

         In our prototype we implement Python Face Recognition Module to train the Face Images. We use KNN algorithm to analyze the nodes in face image then marks the patterns in various images which is taken in different angles. These images get trained as models in python server. We develop a web application as our ground work to mark student’s attendance during Online Exams. We develop with AJAX Api calls java-script functions to get our response and request more responsive. The status of the application and all details of student will be stored and retrieved from MYsql Database Server which is maintained periodically. We implement JDBC connection in java Servlet to access our database. All the requests are sent to the Backend Business Logics which is written in Java Servlets using J2EE technology.

Advantage:

• The advantage of student face verification then student attendance update.
• This can be bridled by face detection and recognition techniques using KNN algorithm.
• The student write their exams in proper way without malpracticing.

Disadvantage:

• The student face recognition take some time.
• Face verification is a process of recognizing and matching faces. The use of biometrics for recognition systems has the aim of increasing human comfort and security in the scope of personal privacy and in a wider scope such as for an agency.